The exercise-conditioned human serum and skeletal muscle cells secretome induce apoptosis in breast cancer cells.

Breast cancer mortality is driven predominantly by metastasis, which affects 20–30% of patients with early-stage disease despite guideline-directed therapies. Because conventional imaging modalities currently lack sensitivity to identify residual disease, molecular-level monitoring must be developed. Circulating tumor DNA (ctDNA) profiling currently enables transformative minimal residual disease (MRD) detection and can quantify tumor burden at low variant allele frequencies. This review provides a comprehensive overview of MRD in breast cancer, including its definition, detection technologies, positivity thresholds, pathophysiology, clinical applications in adjuvant and neoadjuvant settings, ongoing clinical trials, challenges, and future directions. ctDNA-defined MRD has potential as a precision tool for adaptive therapy, and might facilitate post-adjuvant interception, whereby targeted therapies are administered to eradicate micro-metastases before radiographic recurrence. Persistent challenges include MRD assay standardization, subtype-specific MRD thresholds, tumor heterogeneity, and positioning MRD as a potentially valuable tool for precision management in breast cancer.

Aim. The analysis of the dynamics of breast cancer (BC) incidence and mortality among the female population of the Republic of Dagestan for 2014–2022. Materials and methods. The study is based on the analysis of statistical data on the incidence of BC among the female population of the Republic of Dagestan for the period from 2014 to 2022. The analysis uses methods of epidemiological, statistical, comparative analysis and graphical interpretation of data. Results. The average incidence rate of BC over the period from 2014 to 2021 was 17.7 ± 0.3 cases per 100,000 population. The dynamics of incidence over the eight-year period revealed an increase in the discussed indicator by 6.7 %. The standardized incidence rate of breast cancer in 2014 was 41.5 % lower than the all-Russian average (28.4 cases per 100,000 population), and in 2021 it was 38.2 % lower (28.8 cases per 100,000 population). The largest share in the structure of incidence was among patients with stage II of BC. Noteworthy is the tendency to increase from 2019 to 2022 the proportion of patients with stage I from 11.8 to 20.8 % due to a decrease in the proportion of patients with stage III – from 26.1 to 19.6 %. The peak incidence of the discussed pathology in all years of the study was observed in the age group of 55–64 years, with its gradual increase from 30–40 years and decrease after 75 years. The number of deaths due to BC over the analyzed period decreased from 224 cases in 2019 to 163 cases in 2022. At the same time, one-year mortality in 2019 was 15.6 % ( n = 35) of the total number of deaths, in 2020 it was 23.4 % ( n = 43), in 2021 it was 15.6 % ( n = 28), and in 2022 it was 16.6 % ( n = 27). The standardized BC mortality rate of BC in 2021 was 12.8 cases per 100,000 population. The largest number of fatal cases was observed in the age group of 55–69 years, with its gradual increase from 40–45 years and decrease after 70 years, followed by an increase towards 80 years. Conclusion. The conducted population study allowed us to consider in detail the features of the dynamics of BC incidence and mortality rates, population coverage of screening programs and diagnostic methods. All the data obtained and analyzed indicate the need to pay more attention to regional peculiarities, as well as to improve the medical infrastructure of the region.

Abstract Metastasis is the principal cause of mortality in breast cancer, but therapies specifically targeting metastatic mechanisms are scarce. In triple‐negative breast cancer (TNBC), hypoxia within the tumor microenvironment (TME) promotes endothelial dysfunction, increasing vascular permeability and facilitating cancer cell intravasation. This study presents a microfluidic‐based idealized microvascular on‐chip (iMVoC) model utilizing human umbilical vein endothelial cells and TNBC cells (SUM159PTX) to model a hypoxic TME. This model mimicked dynamic flow perfusion, promoting endothelial alignment along the flow direction, while supporting 3D tumor structures exhibiting varying oxygen levels in the tissue compartment. The iMVoC model enabled cell–cell interactions and the exchange of media and nutrients between compartments. Hypoxia was confirmed by increased nuclear translocation of hypoxia inducible factors (HIF)‐1α and HIF‐2α in TNBC cells, indicating hypoxia‐based signaling. Hypoxia‐induced endothelial cell (EC) inflammation was validated through elevated permeability, upregulation of adhesion molecules, and increased reactive oxygen species (ROS) production, suggesting activation of the HIF‐ROS pathway. Enhanced tumor cell intravasation was observed across inflamed endothelium, and cytokine profiling further confirmed EC activation through inflammatory signaling. Application of the protein kinase C delta (PKCδ) inhibitor (PKCδ‐TAT) significantly mitigated these effects, shifting HIF localization from the nucleus to the cytoplasm, reducing ROS production, downregulating inflammatory cytokines, and lowering TNBC intravasation. These findings demonstrate PKCδ as a key mediator linking hypoxia to EC dysfunction and tumor dissemination. Protecting EC barrier integrity emerges as a promising strategy to mitigate hypoxia‐driven TNBC metastasis, with the iMVoC platform offering a valuable tool for testing anti‐cancer therapeutics or drug combinations involving PKCδ‐TAT.

Relevance: Breast cancer (BC) is the most common form of malignant neoplasm among women worldwide. In 2020, 2.3 million new cases and about 685,000 deaths were registered. More than 80% of the cases are women over 50. Developing countries have higher mortality rates. An increase in incidence to 3 million cases by 2040 is forecasted. This study is the first comprehensive 10-year regional analysis of breast cancer incidence, mortality, and stage at detection. The study aimed to analyze the impact of measures implemented in the Almaty region (Kazakh-stan) for early detection and treatment of breast cancer on the dynamics of morbidity, mortality, and stage of detection in 2015-2024. Methods: Assessment of trends and distribution of BC morbidity and mortality rates among the fe-male population of the Almaty region from 2015 to 2024. Statistical reporting forms No. 7, No. 090/U, and data from the regional cancer registry were used. Demographic data were obtained from the official public materials of the Agency for Strategic Planning and Reforms of the Republic of Kazakhstan (Committee on Statistics). The indicators were calculated using standard epidemiologi-cal formulas, direct standardization, and statistical software programs, including Microsoft Excel and SPSS Statistics 23.0. Results: The incidence of BC increased from 34.8 to 42.5 per 100,000, and the standardized rate increased from 34.2 to 39.1. Mortality fluctuated, peaking at 11.6 in 2021, then decreased to 8.5 in 2024. The conditional mortality rate ranged from 20.1% to 35.1%. Early detection at stages I-II in-creased from 74.1% to 89.2% and decreased at stage III from 20.6% to 4.6%. Conclusion: There is a positive trend in early diagnosis and survival in BC in the region. However, the continuing mortality rate and the stable proportion of stage IV indicate the need for further im-provement in the routing and availability of therapy

While parity (number of births) has been extensively studied in relation to breast cancer risk, its age-specific impact on breast cancer mortality remains unclear. We examined how the relationship between parity and breast cancer mortality rates varies across age groups. This retrospective cohort study was based on 894,608 Israeli women born between 1940 and 1960 and followed for 31 years (1990–2020). Employing an age-stratified approach (30–49, 50–64, and 65–80 years), we used Cox regression to examine the associations between number of children (0, 1–2, and ≥ 3) and breast cancer mortality rates, adjusting for sociodemographic variables and calendar year of entry into each age group. Amongst the 30–49 bracket, those with 1–2 children (hazard ratio [HR] = 1.656; 99% confidence interval [CI] 1.349–2.033) or ≥ 3 children (HR = 1.551; 99% CI 1.271–1.893) had a greater breast cancer mortality risk than childless women. In the 50–64 bracket, these differences disappeared after adjustment (HR = 1.071; 99% CI 0.949–1.209 and HR = 0.935; 99% CI 0.830–1.054, respectively). In the 65–80 bracket, those with 1–2 children continued to exhibit an elevated risk (HR = 1.237; 99% CI 1.045–1.466), no significant difference existing between women with ≥ 3 children and childless women (HR = 0.989; 99% CI 0.834–1.173). An age-dependent relationship thus obtained between parity and breast cancer mortality. These findings highlight the need for age-tailored approaches to breast cancer risk assessment and prevention.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12885-025-14993-1.

Introduction: Among breast cancer (BC) survivors, type 2 diabetes (DM) is associated with 40% increased BC mortality. Black and Latina BC survivors have worse BC mortality and are more likely to have DM. We aimed to assess the association of race/ethnicity with adherence to DM self-management behaviors (SMBs), medication beliefs, and control in BC survivors. Methods: We recruited Stages 0-III BC survivors ≥55 years with DM taking ≥1 DM medication. Adherence to DM SMBs and medication beliefs were assessed with validated questionnaires. Hemoglobin A1c (A1c) was measured at recruitment to assess DM control. We evaluated the relationship between race/ethnicity and SMBs, medication beliefs, and DM control with univariate analyses and multivariable regression. Results: Of 194 participants, median age was 67 years (interquartile range [IQR] = 60-72), median A1c was 7.4% (IQR = 6.4-7.9), 27% self-identified as non-Hispanic Black, 41% as non-Hispanic White (NHW), 18% as Latina, and 14% as other/unknown race (not Latina). Race/ethnicity was not significantly associated with adherence to most DM SMBs. In multivariate analyses, non-White race/ethnicity was significantly associated with poorly controlled DM and higher concerns about DM medications, and Black race had a significant association with lower belief of necessity of DM medications compared with NHWs. Conclusions: Race/ethnicity was not associated with DM SMBs in BC survivors, but non-White participants had greater concerns about DM medications and poorer DM control. Future studies exploring differences in DM medication beliefs could inform interventions that can improve DM care among minority BC survivors.

Background: CV mortality (CVM) is a critical competing risk in breast cancer (BC). How therapy type affects CVM by stage and receptor subtype remains incompletely defined. Objective: Compare CVM risk across treatment modalities: chemotherapy alone (C), radiotherapy alone (R), or both (C + R), in BC stages I–IV (2010–2015), stratified by receptor subtype (Luminal A (LA), Luminal B (LB), Her2-enriched, triple negative BC (TNBC)). Methods: Among 282,991 SEER patients (2010–2015), multivariable Cox models (adjusted for age, race, and laterality) and stratified by stage and receptor status estimated adjusted hazard ratios (AHRs) for CVM: C vs R and C + R vs R. Receptor-specific temporal CVM trends are shown in Figure 1. To account for baseline CV comorbidity, 1555 patients from Georgia Cancer Center (2010–2015) were analyzed, adjusting for age, receptor status, stage, hypertension, diabetes, and hyperlipidemia. Results: For the SEER cohort, there was an overall downward trend of CVm across stages 1, 2, and 3. In stage II breast cancer, both LA and LB subtypes showed significant reductions in cardiovascular mortality with C or C + R compared to R. For LA, the AHR was 0.75 (95% CI: 0.59–0.95) for C vs R and 0.75 (95% CI: 0.61–0.93) for C + R vs R. For LB, AHRs were 0.54 (95% CI: 0.33–0.90) for C vs R and 0.44 (95% CI: 0.27–0.74) for C + R vs R. In stage III disease, patients with LA who received C + R had a significantly lower risk (AHR 0.49, 95% CI: 0.35–0.68), as did those with TNBC receiving C + R (AHR 0.33, 95% CI: 0.13–0.82). There were no significant associations for other stage-receptor subgroups (p > 0.05). In the Georgia cohort, neither C nor C + R was associated with lower cardiovascular mortality compared to R (p > 0.05). Conclusions: In stages I–III, Patients selected to receive C (± R) has lower CVM vs R alone in LA (stage II–III), LB (stage II), and TNBC (stage III). CVM did not decline in stage IV (AHR ~1.0), possibly due to expanding use of newer systemic therapies. Temporal trend analysis (Figure 1) suggests broader improvements in CV risk management perhaps due to improvement of cardiac care of BC patients; aligning FDA approval timelines with CVM in BC may further elucidate therapy-related CV risks.

Background: Heart failure (HF) significantly impacts mortality among breast cancer patients. Studies indicate that breast cancer patients aged 65+ with HF have reduced long-term survival as HF poses a greater mortality risk than the cancer itself. Research Question: What are the trends and disparities in HF mortality among US women with breast cancer, and how do these patterns vary by race, age group, and region? Methods: We used the CDC WONDER mortality database to extract age-adjusted mortality rates (AAMRs) per 100,000 US women aged ≥25 years for breast cancer and HF mortality from 1999 to 2020. The Joinpoint Regression Program calculated the average annual percentage change (AAPC) in AAMRs. Results: From 1999 to 2020, HF caused 56,006 deaths in US breast cancer women, showing a decreasing trend with AAPC of -1.16% (95% CI: -1.5 to -0.86). The overall AAMR was 1.94 (1.93 – 1.96). Older adults (≥65 years) showed a significantly higher AAMR [8.88 (8.80 – 8.96)] with an AAPC of -1.17% (-1.45 to -0.93) than younger adults (25-64 years) [0.26 (0.25 – 0.27)] with an AAPC of -1.24% (-2.37 to -0.36). Among the racial and ethnic groups, the highest AAMR was observed in non-Hispanic (NH) Blacks [2.37 (2.31 – 2.43)], followed by NH Whites [2.03 (2.01 – 2.05)], Hispanics [0.98 (0.94 – 1.03)], and NH Asians or Pacific Islanders [0.66 (0.61 – 0.72)]. NH Blacks not only showed the highest mortality burden, but also exhibited a slight increase in AAMRs [AAPC: 0.15% (-0.48 to 0.66)]. NH Asians or Pacific Islanders showed the lowest mortality burden with the highest decline in AAMRs [AAPC: -1.47% (-2.71 to -0.28)], followed by Hispanics or Latinos (-1.27%) and NH Whites (-1.01%). Regionally, the Midwest showed the highest AAMR [2.25 (2.21 – 2.29)], followed by the West [1.95 (1.92 – 1.99)], the Northeast [1.85 (1.81 – 1.88)], and the South [1.81 (1.78 – 1.83)]. The Northeast showed the highest decline in AAMRs [AAPC: -1.48% (-2.06 to -0.98)], and the South showed the lowest [-0.59% (-0.99 to -0.20)]. Women in non-metropolitan areas had higher AAMRs [2.36 (2.32 – 2.41)] than those in metropolitan areas [1.87 (1.85 – 1.88)]. Metro and non-metropolitan areas showed a similar decline in AAMRs over the study period (-1.05% vs. -1.20%). Conclusion: HF mortality in US breast cancer females declined over the past 22 years; however, older adults, NH Blacks, the Midwest, and non-metropolitan areas experienced higher mortality burdens, highlighting the need for targeted interventions.

Background: Anthracycline cardiotoxicity increases breast cancer morbidity and mortality. Glucagon-like peptide-1 receptor agonists (GLP-1RAs) improve cardio-renal outcomes in diabetes, but their effect during anthracycline therapy is unknown. Objectives: To determine whether adding GLP-1RAs to anthracycline-based therapy reduces mortality and major cardio-renal events compared to anthracycline alone. Methods: We conducted a retrospective propensity-score–matched cohort study within the TriNetX Global Collaborative Network (146 health-care organizations) of adult females with breast cancer who received anthracyclines between 2005-2024. Cohort A received ≥1 GLP-1a (n = 1691); Cohort B did not (n = 1691). Outcomes were captured 1-1095 days after index and analyzed with Kaplan–Meier survival and Cox proportional-hazards models. Significance was defined as log-rank P < 0.05 with 95 % confidence intervals (CIs) excluding 1. Cohorts were also matched for medication use [ACE inhibitors, Angiotensin receptor blockers, Sodium glucose transporter -2 reuptake inhibitors, Furosemide, and Spironolactone use]. Results: Mean follow-up: 625 ± 379 vs 728 ± 398 days. GLP-1RA therapy lowered risk of all-cause death (HR 0.19, 0.148-0.247, P < 0.01), cardiac arrest (0.335, 0.12-0.92, P=0.03), atrial fibrillation (0.40, 0.22-0.73, P < 0.01), hospitalization (0.478, 0.39-0.59, P < 0.001), emergency-department visit (0.67, 0.54-0.84, P < 0.001), and progression to stage 4 or 5 chronic kidney disease (eGFR <30 mL/min 1.73 m2; 0.41, 0.29-0.56, P < 0.001) (Table 1). Conclusions: The findings suggest that GLP-1a medications may offer protective benefits beyond glucose control, potentially improving cardiovascular, renal, and overall survival outcomes in breast cancer patients undergoing anthracycline chemotherapy. These real-world data support prospective trials of GLP-1a agents as cardio-renal protective adjuncts in oncology domains. Next steps include conducting prospective clinical trials, exploring mechanisms of action, stratifying patient populations, and expanding research into other cancer types or therapies.

Breast cancer (BC) remains one of the leading causes of cancer-related incidence and mortality globally, with growing evidence suggesting that neighborhood socioeconomic status (n-SES) can influence cancer incidence and survival. However, findings across studies remain inconsistent, and the extent to which n-SES impacts BC outcomes has not been comprehensively understood. A thorough systematic review and meta-analysis were conducted to assess the association of neighborhood socioeconomic status n-SES and BC incidence and mortality risk. A thorough investigation of existing literature was conducted by applying ISI Web of Science, PubMed/MEDLINE, and Scopus until June 08, 2025. Hazard ratios (HR), along with their respective 95% confidence intervals (CI), were integrated utilizing a random-effects model. A validated approach was utilized for assessing the quality of the studies and identification of publication bias, employing the Newcastle-Ottawa Scale for the assessment of quality, conducting subgroup analyses to identify potential sources of heterogeneity, and applying Egger's regression to detect publication bias and perform sensitivity analyses. A total of 17 studies with 779,352 adult subjects were found. The results revealed no association of significance among disadvantaged n-SES and BC incidence risk (HR = 1.19; 95% CI 0.86, 1.65; I2 = 97.7%; p = 0.300; n = 10). However, the subgroup analysis reveals an association of significance among disadvantaged n-SES status and a decreased BC incidence risk in case-control studies (vs. cohort studies) (HR = 0.81; 95% CI 0.72, 0.92; I2 = 0.0%; p = 0.001; n = 2) and for the luminal A cancer sub-type (HR = 0.84; 95% CI 0.73, 0.91; I2 = 29.3%; p = 0.012; n = 3). In contrast, the subgroup analysis revealed an significance association between disadvantaged n-SES status and an increased BC incidence for HER2 + (HR = 1.46; 95% CI 1.14, 1.78; I2 = 0.0%; p = 0.002; n = 2) or the triple-negative cancer sub-type (HR = 1.39; 95% CI 1.19, 1.69; I2 = 41.1%; p < 0.001; n = 4). The results also suggest a significant association between disadvantaged n-SES and a higher BC mortality risk (HR = 1.32; 95% CI 1.16, 1.51; I2 = 76.4%; p < 0.001; n = 8). Sensitivity analyses and an evaluation of publication bias revealed no significant association with BC incidence or mortality risk. Disadvantaged n-SES is identified as a significant element to higher death rates related to BC. By understanding the root causes of inequalities in BC incidence and mortality across different socioeconomic environments, targeted interventions can be tailored to meet the specific needs of vulnerable individuals, ultimately leading toward improved BC outcomes for those at greatest risk.

ACR Appropriateness Criteria® Female Breast Cancer Screening: 2025 Update.

Multivitamin and mineral use: A rapid review of meta-analyses on health outcomes.

Cancer incidence and mortality risk: associations with diet quality scores in a predominantly low-income and Black cohort study.

Global burden of breast cancer attributable to alcohol consumption: a multi-regional observational analysis (1990-2021).

Behavioral factors have been identified as crucial contributors to breast cancer (BC) risk, exhibiting distinct patterns across different regions, sexes, and age demographics. This study aims to analyze the spatiotemporal trends of behavior-related breast cancer (BRBC) at the global, regional, and national scales from 1990 to 2021, while also forecasting its future trajectory up to 2035. Utilizing the Global Burden of Disease 2021 data, we assessed BC mortality and disability-adjusted life years (DALYs) attributable to tobacco, alcohol, diet, and physical inactivity. Age-standardized rates (ASMR/ASDR) and estimated annual percentage change (EAPC) were calculated. Risk contributions, sociodemographic correlations, and population drivers were analyzed. Bayesian modeling projected future trends. Global ASMR and ASDR declined slightly (EAPC: - 1.04/- 0.98), but absolute deaths rose 65.4% (125,701 in 2021). High-SDI regions showed declining rates, contrasting increases in low-middle SDI areas. Western Europe had the highest ASMR (7.43/100,000 in Nauru), while China reported maximal deaths (573,281). Dietary risks dominated BRBC burdens (48.2% deaths), followed by alcohol (21.6%), smoking (18.1%), and inactivity (12.1%). Alcohol-related risks exhibited the widest regional variability. Projections suggest ASMR declines by 2035 (8.16% female; 10% male), yet global deaths may reach 129,175 (female) and 2905 (male). This study first systematically analyzes behavioral risk factors' heterogeneous contributions to BC across socioeconomic levels and cultural contexts, revealing rising burdens, geographic disparities, resource inefficiencies, supporting precision interventions, and strategy translation.

Differences in breast cancer outcomes amongst Caribbean born women in Florida, USA—a population-based analysis

BackgroundBreast cancer is a leading global health challenge, exhibiting significant regional disparities in incidence, mortality, and survival outcomes. This study analyzed the burden of breast cancer in 2022 and projects its future impact by 2050 using GLOBOCAN data.MethodsIncidence and mortality data for breast cancer from 2022 were analyzed across continents, age group, HDI and countries categories. The Average Annual Percent Change (AAPC) from 2018 to 2022 was calculated to project cases and deaths for 2050. Mortality-to-Incidence Ratios (MIR) were computed to assess survival disparities.ResultsIn 2022, Asia accounted for the highest breast cancer incidence (985,817 cases), followed by Europe (557,532) and Northern America (306,307). Africa recorded the highest mortality-to-incidence ratio (MIR) of 0.510, highlighting challenges in early detection and treatment. By 2050, global breast cancer cases are projected to exceed 6 million, with Asia, experiencing the most significant rise (2.0 million cases) followed by Africa (1.118 million cases), followed by. Mortality is expected to rise proportionally, with Asia (484,468) and Africa (390,695 deaths) and bearing the largest burden. The MIR for 2050 shows marked disparities, with Africa (0.35) and Asia (0.25) remaining elevated compared to Europe (0.20) and Northern America (0.13).ConclusionThe projected rise in breast cancer incidence and mortality highlights the urgent need for region-specific interventions. Targeted strategies focusing on early detection, improved access to treatment, and reduction of modifiable risk factors are essential, particularly in transitioning economies where disparities remain stark.

Breast cancer is the most common type of cancer among women, particularly in low and middle-income countries. Breast self-examination is one of the non-invasive methods of breast cancer screening in which a woman looks at her breast for any abnormal findings. Knowledge and practice of breast self-examination increase early diagnosis of breast cancer, which raises the chances for successful treatment and reduction of breast cancer mortality in Ethiopia. Therefore, the objective of this study was to assess knowledge and practice of breast self-examination among female undergraduate students in Jigjiga City. An institution-based cross-sectional study was conducted among 407 randomly selected female students. Data were collected using a structured and pretested questionnaire and entered into SPSS Version 24 for analysis. Descriptive, and bivariate analysis was conducted and variables having p < 0.02 was selected for multivariable logistic regression analyses. The strength of statistical association between breast self-examination and independent variables was measured by an adjusted odds ratio and 95% confidence intervals, and statistical significance was declared at p-value < 0.05. The study revealed that about 28.9% of participants had good knowledge of breast self-examination, while only 17.7% practiced it. Better knowledge was linked to younger age (p = 0.03), discussions about breast self-examination (p = 0.001), knowing someone with breast problems (p = 0.02), and receiving information from colleagues (p = 0.012). Practice was more common among students in higher academic years (p = 0.04), those with a family history of breast cancer (p = 0.011), and those who received information from colleagues (p = 0.001), and health professionals (p = 0.01). The level of knowledge and practice of breast self-examination by undergraduate female students was very low. Age, discussion about breast self-examination, knowing someone diagnosed with breast problems, colleagues as sources of information, academic year, and family history of breast cancer appeared to be significantly influencing the knowledge and practice of breast self-examination. Implementations are needed in addressing young females, making awareness and advocacy campaigns about breast self-examination through different media by stakeholders to increase early diagnosis of breast cancer, which raises the chances for successful treatment in Ethiopia.

ABSTRACT Breast cancer is a common malignant tumor that poses a serious threat to women's health. The incidence and mortality rates of breast cancer have shown an increasing trend worldwide in recent years; therefore, an accurate assessment of breast cancer prognosis is crucial for the development of individualized treatment plans and for the improvement of survival quality of patients. The traditional prognosis assessment of breast cancer mainly depended on doctors' clinical experience and multidisciplinary comprehensive judgment, which lacks unified objective evaluation criteria. This study proposes an innovative cross‐modal contrastive learning model PreGAT based on graph neural networks and attention mechanism. The proposed model can efficiently integrate features from multiple sources of patient data, including clinical features and constructed graph structure features, and significantly improve the performance of the model through a novel contrastive learning loss function. The PreGAT model achieves excellent performance on the public METABRIC dataset with an average accuracy of 92.9% and an AUC value of 0.969. This research provides a promising technique for breast cancer prognosis prediction in clinical practice, which can provide more accurate and reliable decision support for the development of precise treatment programs.